Usage

All the plotting functions use seaborn and matplotlib under the hood.

This means you can show or update the figures accordingly by adding a title. Or if you want to show different label names than present in your truth and prediction values, you can change the labels after plotting.

import audplot
import matplotlib.pyplot as plt
import seaborn as sns


sns.set()  # get prettier plots

truth = ["A", "A", "A", "B", "B", "B", "C", "C", "C"]
prediction = ["A", "A", "B", "B", "C", "C", "A", "A", "C"]
label_aliases = {"A": "c1", "B": "c2", "C": "c3"}

plt.figure(figsize=[2.8, 2.5])
plt.title("Confusion Matrix")
audplot.confusion_matrix(truth, prediction, label_aliases=label_aliases)

plt.tight_layout()

To show multiple graphs in one figure you can specify the axes to draw on.

import audplot
import numpy as np
import matplotlib.pyplot as plt


truth = np.random.randn(100)
prediction = np.random.randn(100)

plot_funcs = [
    audplot.distribution,
    audplot.scatter,
    audplot.series,
]
fig, axs = plt.subplots(1, len(plot_funcs), figsize=[12, 3])
plt.suptitle("Multiple plots in one figure")
for plot_func, ax in zip(plot_funcs, axs):
    plot_func(truth, prediction, ax=ax)

plt.tight_layout()